Machine Learning Reveals Hidden Bias in ERA5 Cloud Heights Over Earth's Third Pole

Abstract Accurate cloud base height (CBH) over the Tibetan Plateau—Earth's Third Pole—is essential for monsoon dynamics, glacial melt, and water security, yet ERA5 systematically underestimates it. Here, we present a two‐step machine learning framework to mitigate this hidden bias. Step 1 refines the ERA5 diagnostic informed by 3 years of ground‐based Lidar observations (October 2021–December 2024), reducing the site‐level MBE from 1.8 to 0.1 km and improving the regional correlation with CALIPSO from 0.25 to 0.40. Step 2 applies an Optuna‐optimized XGBoost model trained on high‐confidence CALIPSO observations, fusing the refined ERA5 data with vertical atmospheric profiles and surface attributes. The final product achieved a test‐set RMSE of 1.87 km ( R 2 = 0.71, MBE = −0.02 km), with seasonal correlations reaching 0.72–0.86 and southern plateau bias reduced to −0.11 km, a 97.9% improvement. This approach enables reliable, long‐term CBH reconstruction, key to better climate modeling and water assessments across High Mountain Asia.